File size: 7,385 Bytes
09f0a1a ca18bc0 09f0a1a ca18bc0 09f0a1a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 |
---
thumbnail: https://github.com/rinnakk/japanese-pretrained-models/blob/master/rinna.png
license: apache-2.0
language:
- ja
- en
tags:
- qwen2
- conversational
base_model:
- Qwen/Qwen2.5-32B
- Qwen/Qwen2.5-32B-Instruct
- rinna/qwen2.5-bakeneko-32b
base_model_relation: merge
pipeline_tag: text-generation
library_name: transformers
---
# `Qwen2.5 Bakeneko 32B Instruct (rinna/qwen2.5-bakeneko-32b-instruct)`
# Overview
This model is an instruction-tuned variant of [rinna/qwen2.5-bakeneko-32b](https://huggingface.co/rinna/qwen2.5-bakeneko-32b), fine-tuned using Chat Vector and Simple Preference Optimization (SimPO).
It adheres to the Qwen2.5 chat format and is designed to deliever superior performance in Japanese language tasks.
| Size | Continual Pre-Training | Instruction Tuning | DeepSeek-R1 Distillation
| :- | :- | :- | :-
| 32B | Qwen2.5 Bakeneko 32B [[HF]](https://huggingface.co/rinna/qwen2.5-bakeneko-32b) | Qwen2.5 Bakeneko 32B Instruct [[HF]](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct)[[AWQ]](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct-awq)[[GGUF]](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct-gguf)[[GPTQ int8]](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct-gptq-int8)[[GPTQ int4]](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct-gptq-int4)| DeepSeek R1 Distill Qwen2.5 Bakeneko 32B [[HF]](https://huggingface.co/rinna/deepseek-r1-distill-qwen2.5-bakeneko-32b)[[AWQ]](https://huggingface.co/rinna/deepseek-r1-distill-qwen2.5-bakeneko-32b-awq)[[GGUF]](https://huggingface.co/rinna/deepseek-r1-distill-qwen2.5-bakeneko-32b-gguf)[[GPTQ int8]](https://huggingface.co/rinna/deepseek-r1-distill-qwen2.5-bakeneko-32b-gptq-int8)[[GPTQ int4]](https://huggingface.co/rinna/deepseek-r1-distill-qwen2.5-bakeneko-32b-gptq-int4)
* **Model architecture**
A 64-layer, 5120-hidden-size transformer-based language model. For a comprehensive understanding of the architecture, please refer to the [Qwen2.5 Technical Report](https://arxiv.org/abs/2412.15115).
* **Training**
This model was developed through a multi-stage training process:
**Model merging.** The base model was augmented with instruction-following capabilities through a Chat Vector addition process.
The Chat Vector was derived by subtracting the parameter vectors of [Qwen/Qwen2.5-32B-Instruct](https://huggingface.co/Qwen/Qwen2.5-32B-Instruct) from [Qwen/Qwen2.5-32B](https://huggingface.co/Qwen/Qwen2.5-32B), as follows.
~~~~text
rinna/qwen2.5-bakeneko-32b + 1.0 * (Qwen/Qwen2.5-32B-Instruct - Qwen/Qwen2.5-32B)
~~~~
During this process, the embedding layer was omitted when performing the subtraction and addition of parameter vectors.
**SimPO** was applied using a subset of the following dataset to further refine the performance of the merged model.
- rinna's internal dataset
* **Contributors**
- [Xinqi Chen](https://huggingface.co/Keely0419)
- [Toshiaki Wakatsuki](https://huggingface.co/t-w)
- [Kei Sawada](https://huggingface.co/keisawada)
---
# Benchmarking
For detailed benchmarking results, please refer to [rinna's LM benchmark page](https://rinnakk.github.io/research/benchmarks/lm/index.html).
---
# How to use the model
~~~python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_id = "rinna/qwen2.5-bakeneko-32b-instruct"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
model_id,
device_map="auto",
torch_dtype=torch.bfloat16,
)
messages = [
{"role": "system", "content": "あなたは誠実で優秀な日本人のアシスタントです。"},
{"role": "user", "content": "ゲーム・小説・アニメに登場するアイテムボックスの特徴と、その原理を詳細に推測してください。"},
]
prompt = tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True
)
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
).to(model.device)
outputs = model.generate(
input_ids,
max_new_tokens=1024,
do_sample=True,
temperature=0.7,
top_k=20,
top_p=0.8,
repetition_penalty=1.05,
)
response = tokenizer.decode(outputs[0][input_ids.shape[-1]:], skip_special_tokens=True)
print(response)
~~~
---
# Tokenization
This model inherits the original [Qwen/Qwen2.5-32B-Instruct](https://huggingface.co/Qwen/Qwen2.5-32B-Instruct) tokenizer.
---
# How to cite
```bibtex
@misc{rinna-qwen2.5-bakeneko-32b-instruct,
title = {rinna/qwen2.5-bakeneko-32b-instruct},
author = {Chen, Xinqi and Wakatsuki, Toshiaki and Sawada, Kei},
url = {https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct}
}
@inproceedings{sawada2024release,
title = {Release of Pre-Trained Models for the {J}apanese Language},
author = {Sawada, Kei and Zhao, Tianyu and Shing, Makoto and Mitsui, Kentaro and Kaga, Akio and Hono, Yukiya and Wakatsuki, Toshiaki and Mitsuda, Koh},
booktitle = {Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)},
month = {5},
year = {2024},
pages = {13898--13905},
url = {https://aclanthology.org/2024.lrec-main.1213},
note = {\url{https://arxiv.org/abs/2404.01657}}
}
```
---
# References
```bibtex
@misc{qwen2.5,
title = {Qwen2.5: A Party of Foundation Models},
url = {https://qwenlm.github.io/blog/qwen2.5/},
author = {Qwen Team},
month = {September},
year = {2024}
}
@article{qwen2,
title = {Qwen2 Technical Report},
author = {An Yang and Baosong Yang and Binyuan Hui and Bo Zheng and Bowen Yu and Chang Zhou and Chengpeng Li and Chengyuan Li and Dayiheng Liu and Fei Huang and Guanting Dong and Haoran Wei and Huan Lin and Jialong Tang and Jialin Wang and Jian Yang and Jianhong Tu and Jianwei Zhang and Jianxin Ma and Jin Xu and Jingren Zhou and Jinze Bai and Jinzheng He and Junyang Lin and Kai Dang and Keming Lu and Keqin Chen and Kexin Yang and Mei Li and Mingfeng Xue and Na Ni and Pei Zhang and Peng Wang and Ru Peng and Rui Men and Ruize Gao and Runji Lin and Shijie Wang and Shuai Bai and Sinan Tan and Tianhang Zhu and Tianhao Li and Tianyu Liu and Wenbin Ge and Xiaodong Deng and Xiaohuan Zhou and Xingzhang Ren and Xinyu Zhang and Xipin Wei and Xuancheng Ren and Yang Fan and Yang Yao and Yichang Zhang and Yu Wan and Yunfei Chu and Yuqiong Liu and Zeyu Cui and Zhenru Zhang and Zhihao Fan},
journal = {arXiv preprint arXiv:2407.10671},
year = {2024}
}
@article{huang2023chat,
title = {Chat Vector: A Simple Approach to Equip LLMs with Instruction Following and Model Alignment in New Languages},
author = {Huang, Shih-Cheng and Li, Pin-Zu and Hsu, Yu-Chi and Chen, Kuang-Ming and Lin, Yu Tung and Hsiao, Shih-Kai and Tzong-Han Tsai, Richard and Lee, Hung-yi},
year = {2023},
url = {https://arxiv.org/abs/2310.04799}
}
@article{meng2024simpo,
title = {SimPO: Simple Preference Optimization with a Reference-Free Reward},
author = {Meng, Yu and Xia, Mengzhou and Chen, Danqi},
journal = {arXiv preprint arXiv:2405.14734},
year = {2024}
}
```
---
# License
[The Apache License, Version 2.0](https://opensource.org/license/apache-2-0) |